Cautionary tales on heating-rate prescriptions in kilonovae

A major ingredient for kilonova lightcurves is the radioactive heating rate and its dependence on the electron fraction and velocity of the ejecta and, in principle, on the nuclear mass formula. Heating-rate formulae commonly used as the basis for kilonova models are, strictly speaking, incorrect for electron fractions other than Y₄ = 0. 04. Here, we introduce new semi-analytical models for kilonovae with better heating rate prescriptions valid for the full parameter space of kilonova velocities and electron fractions. This new prescription produces, on average, dimmer kilonovae at peak that decay more slowly as compared to previously used prescriptions for otherwise identical kilonova physics. We show the dangers of using inappropriate heating rate estimates by simulating realistic observations and inferring the kilonova parameters via a misspecified heating-rate prescription. While providing great fits to the photometry, an incorrect heating-rate prescription fails to recover the input ejecta masses at greater than 5. This bias from an incorrect prescription has disastrous consequences for interpreting kilonovae, their use as additional components in gamma-ray burst afterglows, and understanding their role in cosmic chemical evolution or for multi-messenger constraints on the nuclear equation of state. Given the true heating-rate is uncertain, we estimate there is a 5 and 10\% systematic uncertainty in the measured ejecta masses and velocities, respectively. For lanthanide-rich ejecta, the systematic uncertainty could be as high as 50\%. This systematic uncertainty limits the precision of any measurements that rely on accurate estimates of kilonova ejecta properties.